Hangout for experimental confirmation and demonstration of software, computing, and networking. The exercises don't always work out. The professor is a bumbler and the laboratory assistant is a skanky dufus.

Recent Items
 
Smartcard Trustworthiness
 
Emulating Obsolete Document Formats
 
If Metadata Is the Answer, What Is the Question?
 
Identifying the Objective, not its Location
 
Criteria for Web Application Security
 
Who Can You Trust?
 
Perfecting Secure Coding
 
The Future of Software Tools
 
It's You? Ping You're It!
 
Self-Publishing Boxed Sets

Archives
2004-06-13
2004-06-20
2004-06-27
2004-08-29
2004-09-05
2004-09-12
2004-09-19
2004-10-10
2004-10-24
2004-11-07
2004-11-28
2004-12-05
2004-12-12

When to Optimize, When to Tune

ACM News Service: Think Like a Customer, Use Your Stopwatch.  2004-10-31: This is a highly-pragmatic and practical set of views, especially with regard to where to tune and when to find an algorithmic optimization.

Geoff Koch's 2004-10-15 SD Times article carries a more-careful, progressive presentation of how tools are used and the sensibilities that are applied when searching for improved performance.  There are some interesting tips, including the importance of making sure that optimized code is not unmaintainable.

Randy Camp of Musicmatch uses an approach that is interesting to me because he talks about focusing on the place where the greatest performance loss occurs.  Also, there is a reference implementation created to provide a framework for assessing improvements.  Since I am building a reference implementation for a project, it will be useful to consider how it could be used to optimize for a simple production version based on that design.

The article has some fanciful ideas at the end, including having appropriate algorithms be automatically selected based on conditions at hand.  This is in fact something functional programming is set up to do, and I'll tuck that thought away too.

posted by orcmid at 12/1/2004 06:28:59 PM, rating data by NewsGator Online